Assessing retinal microvasculature as a biomarker for systemic vascular dysfunction in preeclampsia: A case-control study

Introduction

Preeclampsia (PE) is one of the leading contributors to maternal morbidity and mortality worldwide, complicating up to 10% of pregnancies a year.¹ Preeclampsia is characterized by the onset of hypertension at or after 20 weeks of gestation with the presence of proteinuria. Preeclampsia is also commonly characterized by abnormal placentation.² Preeclampsia is associated with other end-organ disorders, such as kidney dysfunction, visual disturbances, and liver dysfunction.³ It has been noted that preeclampsia manifests in two stages. Stage one is characterized by irregular placentation, followed by stage two, which is facilitated by systemic endothelial dysfunction.⁴ Though the exact cause of PE remains elusive, studies have reported the placenta as the primary instigator of the condition.^5,6 Under normal circumstances, the cytotrophoblasts invade and remodel the spiral arteries of the maternal decidua to increase perfusion to the placenta.⁷ Failed or abnormal placentation, as observed in cases of PE, results in utero-placental hypo-perfusion. Decreased blood flow to the placenta may create a low oxygen environment which triggers the release of anti-angiogenic factors such as soluble endoglin (sEng) and soluble fms-like tyrosine kinase-1 (sFlt-1) into the maternal circulation.⁸ These circulating anti-angiogenic factors bind to circulating pro-angiogenic factors, such as vascular endothelial growth factor (VEGF) and placental growth factor (PlGF), reducing their concentration in the circulation^9,10 The imbalance between maternal anti and pro-angiogenic factors mediates downstream effects that result in endothelial dysfunction.¹¹

Ordinarily, pregnancy induces physiological stress that affects every organ system, including the visual system.¹² Pathological changes within the retinal microvasculature may manifest as a result of newly developed disorders in pregnancy, including cases of preeclampsia.¹³ Modifications of the retinal microvasculature in preeclampsia-induced retinopathy are similar to those observed in hypertensive retinopathy.¹⁴ Reported changes include the narrowing of retinal arteries.¹⁵ Optical coherence tomography angiology (OCTA) offers a reproducible, non-invasive method to evaluate changes within the retinal micro-vasculature.¹⁶ An approach that strives to combine the reflective nature of retinal micro-vasculature with maternal macro-vasculature, as assessed by vascular function, as well as biomarkers of endothelial dysfunction, may assist in creating a standardized protocol for early detection of preeclampsia. This may, in turn, aid in early management, subsequently lowering the maternal morbidity and mortality rates associated with preeclampsia. This study, therefore, aimed to assess microvasculature as a biomarker for systemic vascular dysfunction in preeclamptic women from the Eastern Cape, South Africa.

Methods and materials

Results

A total of 108 participants were recruited – 49 preeclamptic women and 59 normotensive women. The questionnaire assessed lifestyle habits and collected family history of diabetes and hypertension. Additionally, obstetric information was collected.

Lifestyle comparisons showed that a higher percentage of preeclamptic women consumed alcohol within the past 12 months compared to normotensive women (22.4% vs 20.3%). A lower percentage of preeclamptic women were smokers (6.1% vs 16.9%). A higher percentage of preeclamptic women had a family history of diabetes (24.5% vs 8.5%) and hypertension (30.6% vs 10.2%). Preeclamptic women had more children (81.6% vs 71.6%), and a higher percentage had been diagnosed with preeclampsia in previous pregnancies (36.7% vs 6.8%) ( Table 1).

Maternal age, gestational age, and anthropometric variables (Height, Weight, BMI, and waist-hip ratio) were similar between the two groups. Preeclamptic women had a higher average parity than normotensive women (1.7 ± 0.17 vs 1.2 ± 0.16, p = 0.02) ( Table 2).

Systolic blood pressure (SBP) and mean arterial pressure (MAP) were higher in the preeclamptic group compared to the normotensive group (126 ± 2 vs 114 ± 2, p = 0.001; 94.39 ± 1.98 vs 87.93 ± 1.22, p = 0.002). Although it was not statistically significant, diastolic blood pressure (DBP) demonstrated increased values in the preeclamptic group (78 ± 2 vs 74 ± 1; p = 0.06). Carotid-femoral pulse wave velocity (cfPWV) was higher in the preeclamptic group (7.63 ± 0.2 vs 7.26 ± 0.2; p = 0.05), while ABI was lower in the preeclamptic group (1.16 ± 0.2 vs 1.23 ± 0.3; p = 0.02). The preeclamptic group showed a lower CRAE (129.28 ± 2.49 vs 150.66 ± 3; p = 0.001), a higher CRVE (255.24 ± 4.6 vs 235.09 ± 4; p = 0.001), and a lower AVR (0.51 ± 0.01 vs 0.65 ± 0.14; p = 0.001) ( Table 3).

Pearson’s correlation analysis was performed to assess the relationship between vascular function and hemodynamic parameters in blood pressure-specific groups. Within the preeclamptic group, we observed that SBP (r = 0.41, p = 0.003) and DBP (r = 0.63, p = 0.001) showed a positive relationship with cfPWV while in the normotensive group, heart rate correlated positively with cfPWV (r = 0.32, p = 0.01) and negatively correlated with AVR (r = −0.28, p = 0.04) ( Table 4).

Pearson's correlation analysis between microvascular and macro-vascular function parameters showed a significant negative relationship between cfPWV and CRAE (r: −0.33, p = 0.02) and AVR (r: −0.44, p = 0.001) respectively. On the other hand, ABI correlated positively with CRAE (r: 0.47, p = 0.001) ( Table 5).

Stepwise linear regression was used to assess how well CRAE, CRVE and AVR may predict cfPWV and ABI. cfPWV and ABI were dependent variables while CRAE, CRVE, and AVR were used as independent variables to generate regression models for systemic vascular function (only significant results were reported). The analysis showed that AVR is a negative, statistically significant predictor of cfPWV (β: −5.67; p = 0.008). The adjusted R² of 0.42 indicates that 42% of the variable can be explained by the model. On the other hand, although there was a positive relationship between ABI and CRAE (β: 0.28; p = 0.005), there was a weak fit in the model ( Table 6).

Discussion

This cross-sectional study assessed the relationship between macro-vascular and microvascular function in preeclamptic women. Our findings showed that women with preeclampsia had increased arterial stiffness and decreased retinal function compared to normotensive women. Our findings showed that arterial stiffening, as assessed by cfPWV, significantly correlated with hemodynamic measurements in the preeclamptic study group. Similarly, retinal microvascular narrowing and stiffening were positively correlated with hemodynamic measurements. Furthermore, we observed a positive association between arterial stiffening and retinal microvascular dysfunction. Increased peripheral arterial disease, as assessed by ABI, was positively associated with retinal arteriolar narrowing. These findings indicate that increasing arterial stiffening may contribute to increasing blood pressure. This relationship is significantly demonstrated in women with preeclampsia. Changes within the microvasculature, assessed by AVR, could predict changes in arterial stiffening. This means that alterations in microvasculature may lead to changes within macrovasculature, which has been shown to relate to blood pressure alterations.

Pulse Wave Velocity (PWV) is the quantification of the speed at which arterial pulses move through the arterial network, with a greater PWV value signifying increased arterial stiffness.²² A prospective study conducted by Phan et al²³ in a Canadian population revealed a similar increase in cfPWV in both normotensive and women at risk of developing preeclampsia at 10 to 13 weeks’ gestation. However, from 20 to 25 weeks, there was an observable difference in cfPWV, with higher levels observed in preeclamptic women, indicating an increase in arterial stiffening. Eastabrook et al²⁴ also observed higher levels of cfPWV in preeclamptic women, compared with normotensive women. Additionally, another study conducted in Uruguay showed that normotensive pregnant women had lower cfPWV compared to preeclamptic women (5.6 ± 0.8 vs 8.2 ± 1.2; P < 0.001).²⁵ These findings are in consensus with those observed in our current study. Peripheral arterial disease is a dominant atherosclerotic state and is commonly linked to impaired arterial function. It reliably predicts cardiovascular disease-related mortality and morbidity.²⁶ Ankle-Brachial index, on the other hand measures the severity of peripheral arterial disease, with normal values ranging between 1.0 and 1.4.²⁷ In the current study, preeclamptic and normotensive women had ABI values that were within normal ranges. However, preeclamptic women had lower ABI, indicative of increased peripheral arterial damage compared to normotensive women.

Previous studies have used retinal vessel calibers to assess cardiovascular risk.²⁸ Narrower retinal arterioles and more dilated retinal venules have been associated with adverse cardiovascular outcomes, such as incidence of stroke and coronary artery disease. We observed that women with preeclampsia had narrower retinal arterioles (smaller CRAE), wider venules (larger CRVE), and smaller AVR, compared to normotensive women. Similar to our findings, a study conducted in the Shanghai First Maternity and Infant Hospital of the Tongji University observed that when compared to normotensive pregnancies, preeclamptic women had narrower CRAE (94 (87, 99) vs 87 (80,94) P < 0.001) and smaller AVR (0.75 (0.71, 0.81) vs 0.72 (0.67, 0.77), P = 0.002).²⁹ Additionally, stepwise linear regression models performed in this study revealed that AVR moderately predicted cfPWV and that CRAE was associated with ABI. A study conducted in the Netherlands with a population of 2434 participants found no associations between arterial stiffness and retinal microvascular calibres.³⁰ These findings are at variance with those of the current study.

From the current study, our findings demonstrate that changes within macro-vascular parameters are associated with changes in blood pressure. This association is more evident in preeclamptic women. Additionally, our findings show that microvascular alterations may predict macro-vascular changes, ultimately influencing blood pressure dynamics. This means that changes within the microvasculature may lead to observable changes in systemic vascular function.

Conclusion

There is evidence of an existing relationship between microvascular and macro-vascular dysfunction in women with preeclampsia. This study showed that retinal vascular changes are associated with arterial stiffness, directly affecting blood pressure variables. This suggests that assessment of retinal vessel calibres may serve as a non-invasive indicator of systemic vascular health in preeclampsia.

Ethics and consent to participate

Ethical clearance was sought from the Walter Sisulu University, Faculty of Health Sciences Research Ethics Committee, Walter Sisulu University (Ethics Approval No: 223/2024). Permission to recruit participants from hospitals was obtained from the Eastern Cape Department of Health and the Frere and Nelson Mandela Academic Hospitals, respectively.

Informed consent statement

The study was conducted in accordance with the Helsinki II declaration. Written informed consent was obtained from participants before their inclusion in this study.